The early design phase of vehicle development is a challenge for NVH because test hardware is not yet available. Surrogate test data may be available and can be combined with best engineering practices to make design decisions, but accurate prediction capability for modeling NVH effects of proposed changes for new vehicles programs is valuable. SEA is a mature CAE technology for NVH development in the automotive industry worldwide, particularly for airborne noise predictions and at higher frequencies (above 200 Hz). Unlike FEA and deterministic modeling techniques, SEA accuracy relies on a good characterization of acoustic absorption, structural damping, and details of flanking paths, particularly leakage. However, an SEA model correlated to test data that isolates the contribution and mechanisms of each of the dominant paths is a resource that can be used for early design phase NVH predictions of noise transfer paths. The main advantages of an SEA model are insensitivity to geometry details (so early design phase noise and contribution path predictions can be performed and remain accurate as shape details evolve) and the ability to quickly perform NVH sensitivity studies for several new vehicle programs based on a small number of correlated template SEA models. This paper illustrates the process of using test to correlate an SEA model that can then generate noise predictions to support future NVH design studies with higher confidence in the prediction accuracy.